Resource utilization among patients with sepsis syndrome

OBJECTIVE: To assess the resource utilization associated with sepsis syndrome in academic medical centers. DESIGN: Prospective cohort study. SETTING: Eight academic, tertiary-care centers. PATIENTS: Stratified random sample of 1,028 adult admissions with sepsis syndrome and all 248,761 other adult admissions between January 1993 and April 1994. The main outcome measures were length of stay (LOS) in total and after onset of sepsis syndrome (post-onset LOS) and total hospital charges. RESULTS: The mean LOS for patients with sepsis was 27.7 ±0.9 days (median, 20 days), with sepsis onset occurring after a mean of 8.1 ± 0.4 days (median, 3 days). For all patients without sepsis, the LOS was 7.2 ± 0.03 days (median, 4 days). In multiple linear regression models, the mean for patients with sepsis syndrome was 18.2 days, which was 11.0 days longer than the mean for all other patients (P < .0001), whereas the mean difference in total charges was $43,000 (both P < .0001). These differences were greater for patients with nosocomial as compared with communityacquired sepsis, although the groups were similar after adjusting for pre-onset LOS. Eight independent correlates of increased postonset LOS and 12 correlates of total charges were identified. CONCLUSIONS: These data quantify the resource utilization associated with. sepsis syndrome, and demonstrate that resource utilization is high in this group. Additional investigation is required to determine how much of the excess post-onset LOS and charges are attributable to sepsis syndrome rather than the underlying medical conditions.